implement simple quantifier heuristics

The quantifier heuristics library is powerful, but often slow. The
unwind conversions are fast, but have serious restrictions. This commits
adds simple quantifier heuristics. They are based on the techniques
developed for quantifier heuristics library, but only search gap guesses
that don't contain free variables. Some very limited experiments suggest
that it is similarly fast as unwind. It can handle all the cases unwind
can handle. However, it supports also a much richer syntax. Moreover, it
can instantiate in addition to universal and existential quantifiers
also unique existential quantifiers, eliminate select and some.
Moreover, it knows about pairs and lists.

fix selftest of quantHeuristicsLib

The recent changes to how context is processed were unluckily
not propagated to the selftests. This fixes the resulting errors.

quantHeuristicsLib: fix handling of pairs, context cleanup

The heuristic for instantiating product variables was broken.
Nested pairs were broken. Pairs like (a, b, c) worked, while ((a,b), c)
did not. Moreover, context heuristics that recurse are now not any more
part of the basic quant parameter. This stops the quant-heuristics
looping sometimes, when the user isn't careful.

Remove trailing whitespace in src/

minor tweaks on quantHeuristicsLib, adept Holfoot to changes with MEM

quantHeuristicLib tuning
- better name for traces
- list_no_len_qp added
- predefined filter functions added

A bit more work on list_qp in order to handle LENGTH. For
LENGTH with large numbers, list_len_qp added.
Experimental support for removing redundant function application on
quantified variables added.

Delete trailing whitespace in src/

bug-fix in Sanity.sml another cleanup of quantHeuristicsLib

extended interface for quantifier heuristics

Now additional context as well as lemmata used for weakening / strengthening can be added to quantifier heuristic parameters.

Added support for context to quantifier heuristics, add sum_qp

cleanup of quantifier heuristics library

- more sensible names for guesses
- quantHeuristicsLib renamed to quantHeuristicsLibBase
- quantHeuristicsLibArgs renamed to quantHeuristicsLibParameters
- new shorter signature defined in quantHeuristicsLib
- adaptation of usage of quantHeuristicsLib

More on quantfier heuristics (I did not add all changes to the first commit)

Remove trailing whitespace in source files.

- reimplementation of quantHeuristicsLib.
- move parts of quantHeuristicsLib to quantHeuristicsTools
- selftest added

Before guesses were justified by hand crafted theorems. There were modified using low level forward proof techniques. This resulted in quantHeuristicsLib being hard to understand and maintain.

This reimplementation exploits recent addition to pairTools about handling paired quantification in order to introduce general predicates for guess theorems. Defining special predicates allows proving general lemmata easily. quantHeuristicsLib can the use the lemmata instead of low level proofs. However, this comes with the additional cost of introducing and eliminating tupled quantifiers.

The new version is easier to understand and maintain. Moreover, it became much simpler to add user specified heuristics. The behavior changed slightly. For example, unnecessary priming of newly introduced variables has been eliminated.

DISCH_ASM_CONV_TAC and DISCH_ASM_CONSEQ_CONV_TAC moved to ConseqConv.

The changes that were discussed during my ARG Lunch talk last Tuesday.

- "quant_heuristic_combine_argument" (qhca) has been renamed to "quant_param" (qp).
- qp for record-types added
- the pair_qp was modified to allow passing in functions to decide which pairs to split and
how to name the new variables
- for easier interaction with the simplifier "QUANT_INST_ss" was added
- examples updated

Some speed improvements. The most noticeable one is that minimising the number of occurences of a quantified variable (which is a very expensive operation) can be turned off now.

Removed trailing whitespace from all .sml and .sig files.

This affects over 900 files and was done using emacs's delete-trailing-whitespace
function in batch mode. Building the system with Poly/ML and Moscow ML seems to
work, so I'm hoping these changes don't break anything. Please complain if
they do!

VARIANT_TAC and VARIANT_CONV added

Now occurences of a variable are tried to be removed before a possible instantiation is searched. For example
"?x. P /\ (f 2 = f x) /\ (Q (f x))" is simplied to "?x. P /\ (f 2 = f x) /\ (Q (f 2))" before an instantiation for x is searched. This enables the heuristics to instantiate x with 2.

Moreover, an additional field was added to qhcas. Now one can specifiy a list of theorems used for the final simplification after an instantiation has been picked.

Cheanup in interface of quantHeuristicsLib, added documentation to source code, clearer handling of stateful heuristics. Heuristics for common datatypes like pairs, lists, option types and nums have been defined in a new library quantHeuristicsArgsLib

QUANT_CONV added

Consolidation of the change, that quantHeuristics can now be restricted to work on just a few instantiations. The examples have been adapted and new examples for the new features added.

added a filter function, that allows to just instantiate variables that are accepted by this filter

some extensions of the interface, interface for caching added

EXT_PURE_QUANT_INSTANTIATE_CONSEQ_CONV and EXT_QUANT_INSTANTIATE_CONSEQ_CONV added

Major extensions and preformance improvements. Notice that the interface to write own heuristics changed.

quantHeuristics is a library that instantiates quantifiers with guesses found by heuristics. Please see quantHeuristicsLib-examples.txt for some documentation and examples.